透過您的圖書館登入
IP:18.119.110.116
  • 學位論文

應用資料探勘技術於SMT線測試誤判率與其影響因子重要程度分析-以H公司為例

指導教授 : 沈國基
若您是本文的作者,可授權文章由華藝線上圖書館中協助推廣。

摘要


近幾年隨著新產品應用高度成長如智慧型手機和平板電腦,台灣的SMT生產技術又開始受到科技產業高度的重視,面對全球突如其來大幅度成長的市場需求,SMT生產線面臨人力與設備的增加無法趕上需求的窘境。在資源不足的情況下生產的效率必須有效提昇,不必要的重工與浪費必須設法改善。設備的使用、人力的計算、治具的製作也都需要透過更精確的計算,才能使生產產能發揮最大的效益。   本研究旨在SMT測試誤測率進行分析,透過資料探勘的技術針對影響誤測率的相關因子進行探討,過程中主要是利用統計軟體Clementine,對PCBA機板ICT測試紀錄進行資料探勘,挖掘關鍵性因子提供工程單位正確之改善方向,減少試誤的次數及摸索的時間,同時可藉由利用軟體已建立的模型,針對新產品所設定參數值進行誤測率的預測提前預知誤測率概估值,使企業除了能夠快速且有效降低測試誤測率更能讓生產企劃部門預先規劃更準確的生產資源,減少因資源規劃錯誤造成產能不足或過剩的情形,提升在同業中之競爭力。

並列摘要


In recent years, with the highly growing application of new products such as smart phones and tablet PCs, Taiwan SMT production technology has begun to attract high technology industry’s attention. Due to the great rise of global market demand for the sudden, SMT production line is facing a trouble that manpower and equipment increase can not catch up with demand. The production must enhance effectively despite the lack of resources. Unnecessary rework and waste have to be improved. The use of equipment, manpower calculations, fixture production also need to be done with more accurate calculation, in order to maximize the benefits of production capacity. This study is aimed at the analysis of SMT’s test error rate, which is through data mining techniques to explore the relevant factors. The process was done with statistical software Clementine, on the PCBA board ICT test records for data mining. Meanwhile, it explored the key factors to provide improvement plans to the depart of engineering correctly. We can also use the model created by software, evaluate the new product set by the parameter values for test error rate. To enhance competitiveness in the same industry, the enterprises would be able to quickly and effectively reduce the test error rate, the production planning department can make resource plans more accurately, and avoid the capacity planning excess or lacking.

並列關鍵字

related factors SMT data mining

參考文獻


1.王堅斌,「使用分類結合法與決策樹判別忠誠顧客」,國立成功大學資訊管理研究所碩士論文,民國92年。
8.林芝儀,「應用資料探勘於信用卡授信決策模式之實證研究」,元智大學工業工程與管理學系,民國91年。
16.許哲榮,「應用影像分割法結合倒傳遞類神經網路於印刷電路板之光學檢驗」,大同大學,民國96年。
27.劉玉婷,「決策樹演算法於LED製程量測站點亮度品質之分析」,國立成功大學,民國96年。
1. Tam, K. Y., and Kiang, M. Y. “Managerial Applications of Neural Networks: The Case of Bank Failure Predictions”, Management Science, pp.926-947, 1992

延伸閱讀